For Digital Photography, it is another story.
How to choose a Digital Camera
Interesting facts of a Digital Videocam's night vision function
Animation - 160 x 120
A collection of digital video COMING ...
Still Picture - 320 x 240
A collection of digital camera taken snapshots COMING ...
How to Choose a Digital Camera (DC)
The display resolutions of different media
| Medium | dpi (minimum required) | How large a 640x480 image display on it? |
| CRT Monitor | 72 | 8.9 in x 6.7 in |
| LCD Monitor | 96 | 6.7 in x 5 in |
| Printing on paper | 300 | 2.1 in x 1.6 in |
For a 2 MPixel DC, the maximum size of the image is 1600 x 1310. If displayed on the LCD the size is 16.7 in x 13.6 in. (Do you have such large LCD monitor?) On the other hand, it seems small if it is printed by a 600 dpi color printer. The size of the printed image by the 600 dpi printer is 2.7 in x 2.2 in. Today most of the ink-based printers is in 1200 dpi.
The size of a 3R photo is 5 in x 3.5 in. The calculated size of this image for 1000 dpi is 16.7MPixel. Today the high-end DC in the market has 5MPixel. Consider if the pixel density double every year. Then it will take two years (2004) to go into the market. This apply to the storage medium too. For example, the largest memory of Sony's Memory Stick is 128MB. It will probably reach 512MB in 2003. Immediately, there is a solution, put a 2.5" 5GB hard disk in the DC. Although there is hard disk based MP3 mobile device, we cannot find this in DC. Then all is for 3R-size, the quality of image is not comparable to the traditional photo prints when the size is as large as 16 in x 20 in. (now you could see the role of scanner)
The principle s of physics will not change with the advance of technology. First, consider the monochromic light, there are five primary aberrations (Seidel aberrations):
| ABERRATIONS | NOTES |
| Spherical | When the aperture and focal length are fixed (find in the cheap DC), the amount of spherical aberration varies with both the object distance and the lens shape. What is the size of your lens? * The Hubble Space Telescope was suffered by spherical aberration in 1990 when first placed in orbit. The prinary mirror was too flat at its periphery by about half a wavelength. Rays from its central region were focusing on the optical axis in front of those from the edges. |
| Coma | The effective focal lengths and therefore the transverse magnifications will differ for rays traversing off-axis regions. Like spherical aberration, coma is dependent on the shape of the lens. What is the shape of the lens of your camera? * A slight tilt of the positive lens under the Sun then you will see the comet shape. |
| Astigmatism | When an object point lies a certain distance from the optical axis, the incident cone of rays strike the lens asymetrically, giving rise to this aberration. What is the size of the object compare to the lens? * Greek a-, meaning NOT, and stigma, meaning SPOT or POINT. |
| Field Curvature | A stigmatic imagery is formed when there is a one-to-one forming of image points of object on the iamge surface. At finite apertures the resulting curved stigmatic image is a manifestation of the primary aberration known as Petzval field curvature. The Petzval surface for a positive lens curves inward toward the object plane, whereas for a negative lens it curves outward away from that plane. In photographic lenses field curvature is most undesirable, since it has the effect of rapidly burring the off-axis image. The solution is to place a negative field flattener lens near the focal plane of a postive lens to nullify the inward curvature. Does the cheap DC has the compound lenses? |
| Distortion | The origin lies in the fact that the transverse magnification is a function of the off-axis image distance. * A thin-lens system will have zero distortion if its optical centre is coincident with the centre of the aperture stop. Therefore, a pinhole camera should has no distortion because the rays connecting conjugate object and image points are straight and pass through the centre of aperture stop. Do you think a cheap DC is a pinhole camera. |
The image formed by the lens falls on the camera's CCD to convert to pixels. The information will be lost due to finite quantization.
Requirements Matrix - fill in your own
| Applications | 300K Pixel | 1 - 2 MPixel | 3 MPixel or above | Optical System | Video Capturing | Storage Medium | Web Cam |
| Web | good | better | too much | good | good | 8 - 16 MB | better |
| Online Photo Album | not enough | good | better | better | better | 32 - 64 MB | no |
| Color Printing | not enough | not enough | good | must | better | 64 or above | no |
About Digital Image Detector
The size of the CCD / CMOS of most of the digital camera is smaller than 35mm film (similar to APS film). For example, the size of the CMOS inside the Sigma SD-9 digital SLR is 13.8mm x 20.7mm. Compare with the size of the 35mm film, 24mm x 36mm, it is 3 times (1.7 x 1.7) smaller. The smaller the image formed the smaller the angle of view and therefore the larger the focal length. For SD-9, the effective focal length of the 15 - 30mm zoom lens is 25.5 - 51mm. The conversion ratio of SD-9 is 1.7x. Below is the table of the angle of view of different kind (size) of camera:
| Wide-angle | Standard | Mid-Tele | |
| SLR 135 Film | 28mm / 75 degree | 50mm / 47 degree | 90mm / 27 degree |
| Bronica RF 645 120 Film | 45mm / 75.1 degree | 65mm / 56.1 degree | 135mm / 29.1 degree |
Note: The size of the 120 film is 56mm x 41.5mm. It is 2.69 times (1.64 x 1.64) of 135 fim.
The effective aperture is defined by focal length over F-number. For SLR 50mm standard lens, the effective aperture is
50mm / F2 = 25mm. For the same effective aperture value, the F-number of the digital camera lens is conversion -ratio times of the original. For example, SD-9 28mm/F-2 is equivalent to 47.6mm/F-3.4. As we know the larger the aperture value, the smaller the depth of field. So a digital camera of small CCD / CMOS size would has a longer focal length and larger F-number than a 135 film taking SLR. Below is the table of the conversion ration of a list of digital camera:
| Conversion Ratio | |
| Canon EOS-1D | 1.3x |
| Nikon D-100 | 1.5x |
| Fujifilm S2 Pro | 1.6x |
| Sigma SD-9 | 1.7x |
Sigma SD-9, like Canon EOS1Ds, is also using CMOS as photodetector. The CMOS in EOS1Ds is in the size of 135 film. On the other hand, Sigma SD-9 uses the Foveon X3 technology, has 3 CMOS layers for each red, green, and blue light while the traditional CCD / CMOS only has one single layer absorbs 25% red, 25% blue, and 50% green. Numerical interpolateion has to be taken before writing the data to the memory for the monolayer CCD/CMOS camera.
Apart from the information loss during the color conversion process, because of the dimension of the CCD / CMOS, the four corners will be darker than the central and there is also internal reflection of light to affect the image quality.
Note: Foveon may has the cross-over problem. Moreover, similar to 3CCD DV, the additional layers would reduce the incoming light intensity affects the operation EV.
Because of the high photosensitivity of CCD / CMOS. It is true that digital camera will has a good performance in low luminous and even lighting / reflection conditions. But for consecutive high-speed photo taking, many digital camera may has the limitation of processing power (numerical) and memory bandwidth. The current trend of technology development suggested that it will take at least two years (2004) for this kind of system to be entered into the main market. Recently, high-speed negative films have been improved a lot. Now we can push the film to ISO 3200 for shooting in high-speed (with 8-10 frames per second) mode. For not pushing, ISO400 is already as good as ISO200 and ISO1600 is as good as ISO800. Below is a table of some DC's ISO range.
| ISO (auto/manual) | |
| Canon PowerShot G2 | 50-400 |
| Canon D60 | 100-1000 |
| Canon 1D | 200-1600 |
| Fujifilm S2 Pro | 100-1600 |
| Nikon Coolpix 4500 | 100-800 |
| Nikon Coolpix 995 | 100-800 |
| Nikon D100 | 200-1600 |
| Sony DSC P9 | 100-400 |
| Sony DSC F707 | 100-400 |
| Kodak DX-3600 | 100 / 200 |
| Kodak Pro Back Plus (16 MPixel) | 100 |
| Kodak Pro Back 645 (16 MPixel) | 100-400 |
| Olymous C-50 Zoom | 80 / 160 / 320 |
| Olymous C-5050 Zoom | 64 / 100 / 200 / 400 |
Remark: May be due to the noise effect only the high-end models have the large ISO values.
Interesting facts of a Digital Videocam's night vision function
Some people think the 0 Lux video capturing of DV is using the same infrared recording technology using in night vision it is wrong. The normal DV does not record the infrared generated from a hot body.
The quality of the image has been dropped after the JPEG compression. Moreover, the contrast is not as good as seen from the DV's LCD.
I have read some reports done by the local computer magazines on the DV night visioning. However, many of them mixed up with the thermo-infrared monitoring. The magaziners took several images of the object covered by different materials. Depends on what the objects are and the material of the cover. One or two shown some recognizable patterns. Then the writers warned the lady be carefully in wearing... Do you think the conclusion is reasonable?
